Institute
Open guidance and self-serve diagnostics
Work from the openly licensed mechanisms catalog (CC BY-SA 4.0), and run a diagnostic yourself for a first reading of readiness.
Commissioned work
When the Institute's open standards and diagnostics are enough, and when to commission Kanav Jain, who wrote this framework, to evaluate a healthcare AI system.
Positioning
Institute vs. commissioned work
The Institute publishes the framework. Kanav Jain applies it to one organization's healthcare AI system under commission.
Where to start
Run it yourself, or commission an evaluation?
Ethotechnics.org is the Institute: open guides and diagnostics you can run yourself. Commissioned safety evaluation for healthcare AI is Kanav Jain's own practice.
Commissioned
Commissioned evaluation of healthcare AI
Four formats: a one-hour safeguards review, a two-week readiness sprint on one workflow, an investor diligence memo in five business days, or a two-to-three-week evaluation against FDA and EU AI Act expectations.
Stay self-serve
Use the Institute when the answer is already in the library
Everything below is free to use and needs no engagement.
- Run a diagnostic and file the readout it returns.
- Find the standards and first check for your kind of system on which systems this is for.
- Cite stable permalinks in memos, contracts, audits, or decision logs.
- Share draft standards internally before committing to delivery changes.
Commissioned work
What the work covers
Whether a healthcare AI system can be stopped, overridden, and reversed, for health plans, healthcare AI companies, and investors.
- Safeguards review (1 hour). A written list of gaps, ranked by the harm they could do.
- Readiness sprint (2 weeks). One workflow mapped end to end: who can halt it, how fast, and who owns each fix.
- Investor diligence (5 business days). The clinical safety, evidence, and regulatory exposure of a healthcare AI deal, in a memo you can cite.
- Clinical AI safety evaluation (2 to 3 weeks). A written evaluation against FDA and EU AI Act expectations, with a summary for your board.
The formats are listed under Engagements on kanav.net, checked October 13, 2026. No prices are published. Most of the work falls under the health coverage and care context.
Boundaries
Engagements stay aligned to published standards
Engagements apply the public standard. They do not rewrite it, and they do not certify a system.
- An engagement cites the same standards and mechanisms published here.
- An evaluation states what Kanav Jain found. Accountability for the system stays with the organization that runs it.
- No page on this site reports on an engagement.
Email kanav@kanav.net. Include the system, the decision it makes, and any diagnostic readout you already have.